DNV-RP-F101 Corroded Pipe Calculator (Single Defect Capacity)

Returns the length correction factor Q = √(1 + 0.31·(ℓ/√(D·t))²), the Part B failure pressure P_f of a single longitudinal metal-loss defect, the safe working pressure P_sw = F₁·F₂·P_f, and the utilization against your operating pressure — the DNV-RP-F101 Part B single-defect method. F₁ (Part B's Modelling Factor), F₂, the tensile strength f_u and everything else that carries safety are user inputs — no DNV factor value is built into the card.

Method last updated 2026-10-04 (calculation changelog) · fixture-verified on every build — most recently 2026-10-29.

Built and fixture-verified by Matthew Norris, P.E. — active P.E. licensure in Arizona, California, Kansas, Missouri, North Carolina, Texas.

DNV-RP-F101 is the other standard answer to the question B31G answers: how much pressure can a corroded pipe still carry? The two methods reach that answer differently, and knowing which one you are running matters more than most engineers assume, because they can disagree by a wide margin on the same defect.

B31G — original and Modified alike — is built on a flow stress derived from yield strength and a Folias bulging factor, and it was calibrated against a body of full-scale tests on older, lower-grade line pipe. RP-F101's failure-pressure equation is built on tensile strength f_u and a different length correction Q, and its full-scale validation tests ran on grades up to X65, with X80 covered by material tests and finite-element analysis only. In industry experience RP-F101 is generally less conservative on long, shallow defects, and the difference is not a rounding error: it is common for a defect that B31G rejects to pass an RP-F101 assessment, and that difference is frequently the whole justification for keeping a line in service rather than cutting it out.

The other structural difference is where the safety lives. B31G folds its conservatism into the method — the parabolic or 0.85 area idealisation, the flow-stress definition — and then applies a safety factor on top. RP-F101 Part B separates the two explicitly: its equation returns the failure pressure P_f, and everything protective is applied afterwards through the Total Usage Factor F = F₁·F₂. P_f is not the recommended practice's mean (best-estimate) capacity: that Sec.2.1 equation carries an extra 1.05 factor, so P_f sits about 5% below it. F₁ is the Modelling Factor Part B itself sets — 0.9 in the edition the worked example follows, and not a project choice; F₂, the Operational Usage Factor, is yours — normally your pipeline's design factor. The card asks for both, from your own copy of the recommended practice, and has no default for either: entering F₁ = F₂ = 1 reads the unfactored Part B failure pressure, which is what you want when you are comparing methods or checking somebody else's assessment, and it is not what you want when you are dispositioning a defect.

What F₂ should be is not a number this tool can supply, but in Part B it is a simple quantity: the Operational Usage Factor, normally taken equal to the pipeline's design factor (Sec.4.2) — the design factor the governing code assigns the line, under B31.4, B31.8 or the owner's specification. It does not carry inspection accuracy. Depth-sizing uncertainty — the standard deviation of the ILI depth call and the partial safety factors built on it — belongs to RP-F101's Part A, the partial-safety-factor route. Part B leaves measurement uncertainty to the assessor, and this card does not model it. F₁, the modelling factor, is not part of the F₂ choice — it comes from the recommended practice itself.

Method

The Part B single-defect method for a longitudinal metal-loss defect under internal pressure only (Sec.4.3.2). First the length correction, which converts the defect's slenderness relative to the pipe's characteristic length √(D·t) into a capacity reduction:

Q = √( 1 + 0.31 · ( ℓ / √( D · t ) )² )

Then the failure pressure of the defected section — Part B's Step 1, with no 1.05 factor (the Sec.2.1 mean-capacity equation is 1.05 × this):

P_f = ( 2 · t · f_u / ( D − t ) ) · ( 1 − d/t ) / ( 1 − (d/t) / Q )

and the safe working pressure and utilization, per DNV-RP-F101 Part B, the allowable-stress approach:

F = F₁ · F₂ (F₁ = Part B Modelling Factor; F₂ = Operational Usage Factor, normally your design factor — both entered by you)

P_sw = F · P_f

utilization = P_op / P_sw → PASS when utilization ≤ 1

Read the failure-pressure equation as three factors. The leading term 2·t·f_u/(D−t) is the burst pressure of sound pipe in the mean-diameter form — note the (D−t) denominator rather than D, which is not the Barlow thin-wall form and gives a slightly higher number. The (1 − d/t) numerator is the fraction of wall the defect leaves behind. The (1 − (d/t)/Q) denominator is where length enters: a very short defect has Q ≈ 1, the two terms nearly cancel, and the capacity approaches sound pipe; a long defect has a large Q, the denominator approaches 1, and the capacity collapses toward the plain (1 − d/t) fraction of the sound-pipe burst — the infinitely-long-defect limit, where the remaining ligament simply carries the hoop load on its own.

Inputs are validated rather than trusted: D and t positive with t < D/2, the defect depth strictly less than the wall (a through-wall defect is a leak, not an assessment), f_u, F₁ and F₂ positive, and the operating pressure non-negative. F₁ has no default: the card computes nothing until you enter it from your own copy of the recommended practice. Entering P_op = 0 puts the card in capacity-only mode: it returns P_f and P_sw and suppresses the verdict, which is what you want when you are establishing what a defect can carry rather than checking it against a known pressure. A measured depth over 85% of the wall (d/t > 0.85) is not accepted by RP-F101 (Sec.4.3.2): the card returns FAIL with that reason and no failure or safe working pressure, whatever the operating pressure. Every accepted run is flagged as applying to an isolated defect only.

Limitations — what this calculator is not

• Single isolated defect only. Real corrosion arrives in colonies, and defects that are close enough to interact carry a combined capacity lower than any of them individually. Assessing interaction — and the axial and circumferential spacing rules that decide when two defects must be treated as one — requires the full RP-F101 procedure, not this equation. Running each pit separately through this card and taking the worst is a known way to get an unconservative answer.
• Longitudinal defects under internal pressure. Circumferentially-oriented metal loss, defects in bends and fittings, and any case where axial load or bending is significant are outside this form. RP-F101 has provisions for combined loading; they are not implemented here.
• Blunt metal loss. Corrosion pitting and general wall loss are what the equation describes. Cracks, crack-like features, gouges, laminations and mechanical damage are not metal loss in this sense and can fail well below the capacity this returns.
• f_u is your input and the answer scales linearly with it. Which tensile strength to use — specified minimum, actual from the material certificate, or a derated value — is set by the assessment route and the governing recommended practice, and it is not a free choice. An assessment run on actual mill tensile values and one run on specified minimum are different assessments with different acceptance criteria.
• F2 carries the project-specific safety and this tool does not select it. In Part B it is the operational usage factor, normally the design factor the governing code (B31.4, B31.8 or the owner's specification) assigns the line. It does not absorb inspection accuracy — that is Part A's partial-safety-factor route. F1 is the recommended practice's own modelling factor, not a project choice; entering F1 = F2 = 1.0 gives the unfactored Part B failure pressure with no protection applied at all — that is a deliberate diagnostic mode, not a disposition.
• Measurement uncertainty is not modelled. The depth you enter is treated as exact. Real ILI depth calls carry a tolerance and a confidence level, and the RP-F101 partial-safety-factor route exists specifically to handle that — this card applies whatever single depth value you give it.
• Measured depths over 85% of the wall are not accepted by RP-F101 (Sec.4.3.2), and the card returns FAIL with no pressure. The limit applies to the measured depth and the card treats your entry as exact, so an ILI call near 85% deserves its sizing tolerance before you enter it. Deep defects are normally repaired rather than assessed.
• It returns a pressure, not a disposition or a remaining life. Corrosion grows; a defect that passes today has a future, and setting a re-inspection interval requires a corrosion rate and a growth assessment that this calculation does not perform.

Quick reference — how defect length changes the answer

The length correction Q is the part of this method engineers find least intuitive, so it is worth seeing what it does across the range. Values below are the equation evaluated for the worked example's NPS 12 × 0.375 in geometry, where √(D·t) = 2.18661 in and d/t = 0.4:

The engineering consequence: for short defects, an error in the length measurement barely moves the capacity, while an error in depth moves it a lot. For long defects the reverse is not true — length has saturated, and depth is again the sensitive input. Depth accuracy is what deserves the inspection budget in both regimes.

Worked example — fixture-verified

An NPS 12 line, 12.75 in OD × 0.375 in wall in API 5L X52 material, with the assessment run on a tensile strength of 66,700 psi. In-line inspection reports an isolated longitudinal metal-loss defect 6 in long with a maximum depth of 0.15 in — 40% of the wall. The assessment enters F1 = 0.9, the Part B modelling factor, and an F2 (design factor) of 0.72, and the line operates at 1,440 psi.

Step by step

1. Characteristic length: √(D·t) = √(12.75 · 0.375) = √4.78125 = 2.18661 in.
2. Normalised defect length: ℓ / √(D·t) = 6 / 2.18661 = 2.74398.
3. Length correction: Q = √(1 + 0.31 · 2.74398²) = √(1 + 0.31 · 7.52941) = √3.33412 = 1.82596.
4. Depth ratio: d/t = 0.15 / 0.375 = 0.4.
5. Sound-pipe term: 2·t·f_u / (D − t) = 2 · 0.375 · 66,700 / (12.75 − 0.375) = 50,025 / 12.375 = 4,042.424 psi.
6. Defect factor: (1 − 0.4) / (1 − 0.4/1.82596) = 0.6 / (1 − 0.21906) = 0.6 / 0.78094 = 0.76831.
7. Failure pressure: P_f = 4,042.424 · 0.76831 = 3,105.827 psi.
8. Total usage factor: F = F₁ · F₂ = 0.9 · 0.72 = 0.648.
9. Safe working pressure: P_sw = F · P_f = 0.648 · 3,105.827 = 2,012.576 psi.
10. Utilization: 1,440 / 2,012.576 = 0.71550 → the line runs at about 72% of the defect's safe working pressure, so the verdict is PASS.

A pass, but with materially less margin than a single-factor reading suggests. The defect factor of 0.768 means this 40%-deep, 6-inch defect has cost the pipe about 23% of its burst capacity — noticeably more than the 40% depth alone would suggest is left, because the length correction is doing real work at ℓ/√(Dt) = 2.74. The 0.7155 utilization is against a safe working pressure that already carries both the 0.9 modelling factor and the 0.72 design factor, so the margin to the unfactored Part B failure pressure is 1,440 psi against 3,105.8 psi — a factor of 2.16 — but the margin to the safe working pressure that actually governs is under 30%. A defect closer to the acceptance line than this one has less room to fold F1 in unnoticed than the old single-factor arithmetic implied. What this result does not include is time: corrosion that produced a 0.15 in defect will keep going, and the number that decides the re-inspection interval is the growth rate, not this utilization.

Worked example 2 — capacity-only mode, reading the unfactored failure pressure

The same pipe and the same defect, run with both F1 and F2 entered as 1.0 and no operating pressure entered. This is the diagnostic mode: it strips every protective factor out — including the recommended practice's own modelling factor — and reports what the Part B failure-pressure equation alone says the defected section can carry.

Step by step

1. Q, the depth ratio and the failure pressure are unchanged — none of them depends on F1, F2 or operating pressure: P_f = 3,105.827 psi.
2. With F1 and F2 both entered as 1.0, the total usage factor F = 1.0 and the safe working pressure equals the failure pressure: P_sw = 1.0 · 3,105.827 = 3,105.827 psi.
3. No operating pressure was entered, so utilization is suppressed and the card returns no verdict — status is COMPUTED rather than PASS or FAIL.
4. Comparing the two runs: the main example's F = F1·F2 = 0.648 is discarding 1,093.25 psi, or about 35.2% of the unfactored failure pressure. That is the entire safety margin of the assessment, visible as a single number — and more than a third of it (the F1 = 0.9 share) is the standard's own modelling factor, not a project choice at all.

Capacity-only mode exists for three jobs, and dispositioning a defect is not one of them. It lets you compare methods on a like-for-like basis — RP-F101's unfactored Part B failure pressure against a B31G failure pressure, both before any safety factor. (It is still about 5% below RP-F101's own mean-capacity equation, which carries a 1.05 factor.) It lets you check somebody else's assessment by separating the equation from the factors they chose. And it lets you see how much of your answer is physics and how much is conservatism, which is the question that comes up whenever a defect is close to the line and somebody asks whether the assessment is being unfair to it. Never quote a capacity-only number as a safe pressure — and remember that reaching it takes F1 = 1 as well as F2 = 1, not just a usage-factor entry of 1.0.

Fixture case “capacity-only mode (Pop=0), F1=F2=1 entered reads unfactored Pf (key Pcap)” (tolerance 0.001) — locked in the same release gate as the example above.

Additional verified cases in this fixture

input:  {"D":12.75,"t":0.375,"d":0.15,"L":6,"fu":66700,"usageFactor":0.72,"Pop":1440}
expect: {"Pcap":3105.827,"F1":1,"F2":0.72,"F":0.72,"Psw":2236.196,"utilization":0.64395}
tol:    0.001

input:  {"D":12.75,"t":0.375,"d":0.31875,"L":6,"fu":66700,"F1":0.9,"F2":0.72,"Pop":700}
expect: {"Q":1.82596,"Pcap":1134.47,"F":0.648,"Psw":735.137,"utilization":0.9522}
tol:    0.001

input:  {"D":12.75,"t":0.375,"d":0.33,"L":6,"fu":66700,"F1":0.9,"F2":0.72,"Pop":0}
expect: {"Q":1.82596,"F":0.648}
tol:    0.001

Sources & citations

• DNV-RP-F101, Corroded Pipelines (September 2019, amended September 2025), Part B (allowable-stress approach), Sec.4.2–4.3.2 — failure pressure and safe working pressure of a single longitudinal metal-loss defect under internal pressure, the 85% measured-depth acceptance limit, including the length correction factor Q and the F = F1·F2 usage-factor split. F1 is Part B's modelling factor (0.9 in the edition the worked example follows); it, F2, the tensile strength basis and every other factor are user-supplied from the governing revision.
• ASME B31G, Manual for Determining the Remaining Strength of Corroded Pipelines — the alternative method, built on a yield-derived flow stress and the Folias factor; see the B31G calculator for the same defect assessed that way.
• API 5L, Line Pipe — source of the specified minimum tensile strength; whether the assessment uses specified minimum or actual certified values is set by the assessment route, not by this tool.
• ASME B31.8S, Managing System Integrity of Gas Pipelines — where a remaining-strength assessment sits within an integrity management programme, and the basis for re-inspection intervals that this calculation does not set.
• No usage factor, safety factor or material value is embedded — F1, F2, fu and everything else are entered by the user from the governing revision of DNV-RP-F101. The equation form, including its 0.31 length-correction coefficient and the 0.85 depth limit, is in the code.

Per the source & citation policy, allowable-stress and factor table values are inputs — typed from your governing edition, or autofilled on paid tiers from the embedded datasets enumerated there, always cited by standard, edition, table and line. Where a page does reproduce specific ASME data (the B16.5 ratings, the quick-reference tables), it states the source table and conditions inline.

FAQ

RP-F101 or B31G — which should I use?

Whichever your integrity procedure and the governing regulation specify; this is not usually a free choice. Where it is, the practical differences are these. B31G (original and Modified) is built on a yield-derived flow stress and the Folias bulging factor, is embedded in US pipeline regulation and long-standing operator practice, and is generally the more conservative of the two — markedly so on long, shallow defects. RP-F101 is built on tensile strength with a different length correction, was validated by full-scale tests on grades up to X65 (X80 by material tests and finite-element analysis only), and in Part B separates the failure pressure from the safety factors applied to it. Running both on the same defect is a reasonable diagnostic; picking whichever one gives the answer you want after the fact is not. One operational note: because the two methods disagree most on long shallow corrosion — the commonest ILI finding on aging lines — a fleet-level choice between them materially changes how many anomalies a dig program must address. That is a decision for the integrity management plan, made once and documented, not remade defect by defect.

What should I enter for F1 and F2?

F1 is not really a choice: it is the Modelling Factor DNV-RP-F101 Part B itself sets — 0.9 in the edition the worked example follows — so take it from your own copy of the governing revision; the card has no default and computes nothing until it is entered. F2, the Operational Usage Factor, is what the recommended practice leaves to the assessment, and Part B keeps it simple: Sec.4.2 says it is normally taken as equal to the design factor. In practice that is the design factor the governing code — B31.4, B31.8 or the owner's specification — assigns the line for its location class and fluid. Inspection accuracy does not go into F2. The standard deviation of the depth call, and the partial safety factors built on it, belong to RP-F101's Part A (the partial-safety-factor route); Part B leaves measurement uncertainty to the assessor, and this card does not model it. If a defect is poorly characterised, the remedy is a Part A assessment or a conservatively sized depth, not an adjusted F2. Set F1 = F2 = 1.0 and you are reading the unfactored Part B failure pressure with nothing protecting you.

Why does the sound-pipe term use (D − t) instead of D?

Because it is the mean-diameter form of the burst expression rather than the thin-wall Barlow form. Barlow uses the outside diameter and is the conventional design equation; the mean-diameter form is closer to the actual stress distribution through the wall and is what the RP-F101 failure-pressure equation is built on. For an NPS 12 × 0.375 in pipe the difference is about 3% — 4,042 psi versus 3,923 psi on the leading term — which is small but not negligible when you are comparing this method's output against a B31G number computed on D. If you are cross-checking the two methods by hand, this is one of the two places the arithmetic will refuse to line up; the flow-stress definition is the other. The audit habit that follows: when reconciling this card against a spreadsheet or legacy calculation, check the diameter convention FIRST, before questioning inputs — a 3% systematic offset on the leading term is the signature of a D-versus-(D − t) mismatch, and it has burned more cross-checks than any data error. State the convention on the calculation sheet whenever a result will be compared across methods.

My corrosion is a cluster of pits, not one defect. Can I use this?

Not directly, and this is the most consequential limitation on the page. Defects that lie close enough to each other interact: the combined feature behaves as one larger defect with a lower capacity than any of its parts assessed alone. RP-F101 has explicit interaction rules — axial and circumferential spacing criteria that decide when adjacent metal loss must be combined, and a procedure for assessing the composite profile — and none of that is implemented here. Assessing each pit separately and taking the worst result will produce a number that is too high. Use the full procedure, or a river-bottom profile assessment, for anything that is not a genuinely isolated defect. As a fast triage before invoking the full interaction procedure: if the axial spacing between two features exceeds the interaction limits by a wide margin, they are honestly independent and this card serves; if they are anywhere near the limits, combine first and assess the composite — the combination is always the conservative reading. ILI vendors will supply clustered calls on request, and asking for clustering in the deliverable is cheaper than re-deriving it from the raw feature list.

Does a passing result mean I can leave the defect alone?

It means the defect meets the acceptance criterion at the pressure and with the factors you entered, today. Two things it does not tell you. It does not set a re-inspection interval: corrosion grows, and the interval comes from the growth rate and the margin you have, which means a defect passing at 0.64 utilization and one passing at 0.98 need very different responses even though both say PASS. And it says nothing about the mechanism — a defect under active corrosion with a failed coating and no cathodic protection is a different proposition from a dormant one, whatever the current utilization is. Remaining strength is an input to an integrity decision, not the decision. The disposition that ages well records three numbers alongside the verdict: the utilization today, the assumed growth rate with its source, and the date the utilization reaches the acceptance line at that rate — because that date, not the pass, is the actual decision the assessment feeds. A file of PASS stamps without growth projections is an inspection scheduling problem wearing a compliance costume.

Can I use this on a defect in a bend, a fitting or a weld?

No. The capacity form assumes straight pipe, an isolated longitudinal defect in the pipe body, and internal pressure as the only significant load. Bends and fittings have different geometry and stress distributions; welds have different material properties and their own acceptance rules; and any location where axial load or bending is significant needs the combined-loading provisions of the full recommended practice rather than this equation. Metal loss found in those locations is normally taken to a fitness-for-service assessment with the actual geometry and loading modelled, not screened with a pipe-body equation. One clarification that saves unnecessary escalations: 'near a weld' and 'in a weld' are different findings. Body metal loss that merely sits some distance from a girth weld, outside the weld and its heat-affected zone, is still a pipe-body defect and assesses normally; loss that touches the weld cap or HAZ is what changes category. Read the ILI feature's weld-interaction call, and when the report is ambiguous, the field examination settles it before the assessment method is chosen.

Related calculators

• B31G Corroded Pipe Remaining Strength Calculator (Level 1) — The yield-based alternative on the same defect — usually the more conservative
• Pipeline MAOP Calculator — Maximum Allowable Operating Pressure (ASME B31.8 / B31.4) — The operating pressure the defect capacity is checked against
• Pipeline Dent Depth Screening Calculator (Plain Dents, %OD) — A dent containing metal loss needs both assessments, and neither alone
• Pipeline Hydrostatic Test Pressure Calculator (ASME B31.4 / B31.8) — Re-testing a corroded line: assess remaining strength before setting the test
• API RP 14E Erosional Velocity Calculator (Line Sizing) — If the metal loss is velocity-driven, check the erosional limit